facebook pixel image Leveraging Big Data Analysis for Insurance Fraud

Leveraging Big Data Analysis for Insurance Fraud

Those who work with insurance claims or in an adjacent field understand that fraud is increasingly prevalent, but detecting it is becoming more difficult.

Figures driving change

Insurance fraud costs carriers roughly $80 billion per year because current methods are outdated, according to WNS Decision Point. While a sizable number of claims—between 5 and 10 percent—are classified as fraudulent every year, it’s likely that fraud occurs more often than that.

Despite an industry-wide recognition that fraud is a major problem, rates continue to rise. Actuarial and consulting firm Milliman found that suspected insurance fraud rose almost 27 percent between 2010 and 2012, the last years data was analyzed. The increase in claims represents an escalation to respective firms in time spent per case, cash funneled toward the investigation and ultimately funds paid out to fraudulent cases that weren’t caught.

A large part of the reason behind this trend of outdated methods being used is the overwhelming amount of data that must be analyzed in a modern fraud investigation.

Unstructured data, social media and insurance claim fraud

According to Fleming, nearly 80 percent all of data insurers receive is unstructured and a good portion of that information is coming from social media.

Yes, the very websites we use to put up photos chronicling our latest vacation are actually gold mines for brokers and claims investigators. Unfortunately, they don’t necessarily have all the proper tools to uncover the most valuable nuggets of information, and this results in them seeing less than 20 percent of the data available to them from their company.

With over 1.5 billion active Facebook users in the world, social media has become a trove of data, but understanding it is a time-consuming process, according to Milliman. People use pseudonyms to conceal their true identity, or hide in a sea of anonymity by using common names. But, according to the source, those investigator who perform their due diligence reap the greatest rewards. For instance, Milliman reported that one fraud investigator found an image of a woman wearing her “lost” wedding ring in a photo posted to social media.

But it’s not just social media that’s complicating insurance fraud investigation. Logs and adjuster’s notes are sprawling and comprehensive. While they could prove valuable in identifying connections and links between different cases, companies still deploying manual methods to comb through big data will be unable to truly comprehend the contents of those documents.

Flipping the script on insurance fraud detection

If investigators want to use present-day evidence to detect fraud, isn’t it time they started using a modern tool to do so?

Fleming reported that 4 in 5 insurance firms are behind their competitors when it comes to implementation of big data analytics tools. In a day and age where fraudulent claims are investigated through the effective use of structured and unstructured data, organizations without the tools to sift through vast sums of information will ultimately lose money in the long run because they won’t be able to link together the clues. According to WNS, undetected fraudulent claims can reduce a carrier’s true revenue by around 10 percent each year.

Visallo gives brokers the tools to adapt to the unprecedented rise of the four V’s: volume, velocity, variety and veracity of data. As a big data analytics and investigation tool, Visallo allows users to search, aggregate, and uncover hidden connections within internal and external data in many formats via the use of graphs, link charts, maps and other visualizations.

Additionally, instead of viewing cases in isolation, insurance claims investigators can work in real-time to share valuable insights they’ve found with other team members without sacrificing the security of their data. Administrators can set privilege levels ahead of time to dictate what information can be seen, added or removed by each user.

Leveraging a big data investigative system like Visallo gives insurance carriers a tool to uncover fraudulent claims that would be missed via manual investigation by allowing investigators to deeply explore data and perform link analysis. By giving employees tools that are geared to improve the speed and accuracy of investigation, they can take on more cases without a drop in performance, according to WNS.

It’s time to put an end to the complexity of leveraging big data to combat insurance fraud. It’s time to try Visallo.